Automation & AI

Intelligent Funnel Optimization

Sales contact forms admitted spam, students and fake job titles, inflating MQL volumes and eroding trust. I standardized job data across portfolios, implemented lead‑to‑account matching, introduced layered quality controls (email verification, pattern‑blocking, automated rejection) and added an AI‑assisted validation layer for ambiguous cases. Junk MQLs dropped by 30 %, SQLs increased 20 % and the MQL‑to‑SQL conversion rate doubled.

The Problem

Leads from our contact sales forms were entering the funnel with inconsistent job data, fake contact information, and no quality controls at intake. The system routinely qualified students, competitors, spam, and invalid records as MQLs, flooding sales with noise and obscuring true funnel performance.

I led the effort to build an intelligent qualification infrastructure that enforced quality at every stage. I designed job role standardization models (replacing legacy job title collection), implemented lead-to-account matching workflows to gain account context before CRM ingestion, deployed automated rejection logic across form submissions and email validation, and built an AI-assisted detection layer for edge cases that rules alone couldn't catch.

The impact was immediate: junk MQLs dropped 30%, SQLs increased 20%, and MQL-to-SQL conversion doubled because sales could trust the data and act on accurate signals. This project demonstrates how combining structured governance with intelligent automation can transform a noisy funnel into a reliable qualification engine.

Every contact form submission became a potential MQL regardless of job role validity, account fit, or contact authenticity.

The funnel was optimized for volume, not quality.

The Solution

Standardize Job Data Across Portfolios

Mapped thousands of inconsistent job titles to clean, governed role categories. This stabilized persona targeting, improved scoring and lifecycle accuracy, and ensured consistent segmentation across all product lines.

Implement Lead-to-Account Matching

Designed a HubSpot-based L2A matching workflow that identified key firmographic attributes before records entered Salesforce. This anchored qualification in account context, prevented low-value account creation, and ensured routing decisions were made with accurate business intelligence.

Deploy Multi-Layered Quality Automation

Introduced automated quality controls across the intake process: ESP feedback loops to quarantine invalid emails, dynamic form scripts to block junk patterns (students, non-business domains, throwaway addresses), and automated rejection logic to prevent unqualified records from ever reaching MQL status.

Build AI-Assisted Junk Detection

Created an AI validation layer using Zapier and structured prompts to evaluate ambiguous submissions. The agent identified fake names, student records, and fraudulent entries, capturing edge cases that rule-based systems couldn't detect.

I moved quality control upstream, from post-submission cleanup to intake-level enforcement.

The funnel became a filter that protected the pipeline instead of a gate that let everything through.